1
00:00:01,830 --> 00:00:04,157
Hi, folks. And welcome to Module 13.

2
00:00:04,157 --> 00:00:06,120
In this module, we will start

3
00:00:06,120 --> 00:00:08,790
our discussion of logistic regression.

4
00:00:08,790 --> 00:00:10,830
Logistic regression is appropriate

5
00:00:10,830 --> 00:00:14,190
when we have a binary
or dichotomous outcome.

6
00:00:14,190 --> 00:00:16,920
This is an incredibly
useful statistical technique

7
00:00:16,920 --> 00:00:18,420
in public health practice

8
00:00:18,420 --> 00:00:21,030
because we often have binary outcomes.

9
00:00:21,030 --> 00:00:23,820
Things like, is the person dead or alive?

10
00:00:23,820 --> 00:00:27,690
Recovered or not recovered?
Diseased or not diseased?

11
00:00:27,690 --> 00:00:31,350
All of these can be modeled
using logistic regression.

12
00:00:31,350 --> 00:00:34,680
We became familiar with these
types of outcomes in Epi 1,

13
00:00:34,680 --> 00:00:37,530
when we worked a lot
with two-by-two tables.

14
00:00:37,530 --> 00:00:39,540
But when we have a two-by-two table,

15
00:00:39,540 --> 00:00:43,080
we only have one binary predictor as well.

16
00:00:43,080 --> 00:00:44,970
Well, what logistic regression does

17
00:00:44,970 --> 00:00:47,490
is allow us to model these binary outcomes

18
00:00:47,490 --> 00:00:49,290
with multiple predictors,

19
00:00:49,290 --> 00:00:52,410
just like we did with
multiple linear regression,

20
00:00:52,410 --> 00:00:54,630
so we can start
controlling for confounding

21
00:00:54,630 --> 00:00:58,530
and look at effect
modification within our models.

22
00:00:58,530 --> 00:01:00,120
In this module, we're going to stick

23
00:01:00,120 --> 00:01:04,110
to just one predictor, which
is simple logistic regression.

24
00:01:04,110 --> 00:01:06,960
And the reason is that in
order to create an equation

25
00:01:06,960 --> 00:01:08,820
for our logistic regression models,

26
00:01:08,820 --> 00:01:10,650
there's some algebra involved.

27
00:01:10,650 --> 00:01:12,660
We wanna make sure we feel solid with that

28
00:01:12,660 --> 00:01:14,700
before moving forward.

29
00:01:14,700 --> 00:01:16,350
In the next module, we'll go ahead

30
00:01:16,350 --> 00:01:18,780
and get into multiple logistic regression,

31
00:01:18,780 --> 00:01:21,150
where we can start to look
at multiple confounders,

32
00:01:21,150 --> 00:01:23,430
effect modification, and all the things

33
00:01:23,430 --> 00:01:26,220
you already learned in
multiple linear regression.

34
00:01:26,220 --> 00:01:28,920
Let me know if there are any
questions, and take care.